lucataco/omnigen2

OmniGen2: a powerful and efficient unified multimodal model

Capabilities

1:14:33:416:99:162:33:2Reference ImagesNegative PromptSeed

Cost

Community model (estimated from hardware time)

Input Parameters

Name	Type	Description	Default	Constraints
`image`*	string(uri)	Input image to edit	`—`	—
`cfg_range_end`	number	CFG range end	`1`	min: 0, max: 1
`cfg_range_start`	number	CFG range start	`0`	min: 0, max: 1
`height`	integer	Height of output image	`1024`	min: 256, max: 1024
`image_2`	string(uri)	Optional second input image for multi-image operations	`—`	—
`image_3`	string(uri)	Optional third input image for multi-image operations	`—`	—
`image_guidance_scale`	number	Guidance scale for input image. Higher values increase consistency with input image	`2`	min: 1, max: 3
`max_input_image_side_length`	integer	Maximum input image side length	`2048`	min: 256, max: 2048
`max_pixels`	integer	Maximum number of pixels in output	`1048576`	min: 65536, max: 2359296
`negative_prompt`	string	Negative prompt to guide what should not be in the image	`"(((deformed))), blurry, over saturation, bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), fused fingers, messy drawing, broken legs censor, censored, censor_bar"`	—
`num_inference_steps`	integer	Number of denoising steps	`50`	min: 20, max: 100
`prompt`	string	Text prompt describing the desired image edit	`"Make the person smile"`	—
`scheduler`	string	Scheduler to use	`"euler"`	eulerdpmsolver
`seed`	integer	Random seed. Set to -1 for random seed	`-1`	—
`text_guidance_scale`	number	Guidance scale for text prompt	`5`	min: 1, max: 8
`width`	integer	Width of output image	`1024`	min: 256, max: 1024

imagerequiredstring

Input image to edit

cfg_range_endnumber

CFG range end

Default: 1min: 0, max: 1

cfg_range_startnumber

CFG range start

Default: 0min: 0, max: 1

heightinteger

Height of output image

Default: 1024min: 256, max: 1024

image_2string

Optional second input image for multi-image operations

image_3string

Optional third input image for multi-image operations

image_guidance_scalenumber

Guidance scale for input image. Higher values increase consistency with input image

Default: 2min: 1, max: 3

max_input_image_side_lengthinteger

Maximum input image side length

Default: 2048min: 256, max: 2048

max_pixelsinteger

Maximum number of pixels in output

Default: 1048576min: 65536, max: 2359296

negative_promptstring

Negative prompt to guide what should not be in the image

Default:

"(((deformed))), blurry, over saturation, bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), fused fingers, messy drawing, broken legs censor, censored, censor_bar"

num_inference_stepsinteger

Number of denoising steps

Default: 50min: 20, max: 100

promptstring

Text prompt describing the desired image edit

Default: "Make the person smile"

schedulerstring

Scheduler to use

Default: "euler"

eulerdpmsolver

seedinteger

Random seed. Set to -1 for random seed

Default: -1

text_guidance_scalenumber

Guidance scale for text prompt

Default: 5min: 1, max: 8

widthinteger

Width of output image

Default: 1024min: 256, max: 1024

Version: 5b9ea1d0821aUpdated: 7/25/20262.2K runs